From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: david@fromorbit.com, darrick.wong@oracle.com
Cc: linux-fsdevel@vger.kernel.org, xfs@oss.sgi.com,
Dave Chinner <dchinner@redhat.com>
Subject: [PATCH 15/58] xfs: rmap btree requires more reserved free space
Date: Tue, 06 Oct 2015 21:56:25 -0700 [thread overview]
Message-ID: <20151007045625.30457.21863.stgit@birch.djwong.org> (raw)
In-Reply-To: <20151007045443.30457.47038.stgit@birch.djwong.org>
>From : Dave Chinner <dchinner@redhat.com>
The rmap btree is allocated from the AGFL, which means we have to
ensure ENOSPC is reported to userspace before we run out of free
space in each AG. The last allocation in an AG can cause a full
height rmap btree split, and that means we have to reserve at least
this many blocks *in each AG* to be placed on the AGFL at ENOSPC.
Update the various space calculation functiosn to handle this.
Also, because the macros are now executing conditional code and are called quite
frequently, convert them to functions that initialise varaibles in the struct
xfs_mount, use the new variables everywhere and document the calculations
better.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
---
fs/xfs/libxfs/xfs_alloc.c | 69 +++++++++++++++++++++++++++++++++++++++++++++
fs/xfs/libxfs/xfs_alloc.h | 41 +++------------------------
fs/xfs/libxfs/xfs_bmap.c | 2 +
fs/xfs/libxfs/xfs_sb.c | 2 +
fs/xfs/xfs_discard.c | 2 +
fs/xfs/xfs_fsops.c | 4 +--
fs/xfs/xfs_mount.c | 2 +
fs/xfs/xfs_mount.h | 2 +
fs/xfs/xfs_super.c | 2 +
9 files changed, 84 insertions(+), 42 deletions(-)
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index 3c55fa7..6fd9d3e 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -62,6 +62,72 @@ xfs_prealloc_blocks(
}
/*
+ * In order to avoid ENOSPC-related deadlock caused by out-of-order locking of
+ * AGF buffer (PV 947395), we place constraints on the relationship among actual
+ * allocations for data blocks, freelist blocks, and potential file data bmap
+ * btree blocks. However, these restrictions may result in no actual space
+ * allocated for a delayed extent, for example, a data block in a certain AG is
+ * allocated but there is no additional block for the additional bmap btree
+ * block due to a split of the bmap btree of the file. The result of this may
+ * lead to an infinite loop when the file gets flushed to disk and all delayed
+ * extents need to be actually allocated. To get around this, we explicitly set
+ * aside a few blocks which will not be reserved in delayed allocation.
+ *
+ * The minimum number of needed freelist blocks is 4 fsbs _per AG_ when we are
+ * not using rmap btrees a potential split of file's bmap btree requires 1 fsb,
+ * so we set the number of set-aside blocks to 4 + 4*agcount when not using rmap
+ * btrees.
+ *
+ * When rmap btrees are active, we have to consider that using the last block in
+ * the AG can cause a full height rmap btree split and we need enough blocks on
+ * the AGFL to be able to handle this. That means we have, in addition to the
+ * above consideration, another (2 * mp->m_ag_levels) - 1 blocks required to be
+ * available to the free list.
+ */
+unsigned int
+xfs_alloc_set_aside(
+ struct xfs_mount *mp)
+{
+ unsigned int blocks;
+
+ blocks = 4 + (mp->m_sb.sb_agcount * XFS_ALLOC_AGFL_RESERVE);
+ if (!xfs_sb_version_hasrmapbt(&mp->m_sb))
+ return blocks;
+ return blocks + (mp->m_sb.sb_agcount * (2 * mp->m_ag_maxlevels) - 1);
+}
+
+/*
+ * When deciding how much space to allocate out of an AG, we limit the
+ * allocation maximum size to the size the AG. However, we cannot use all the
+ * blocks in the AG - some are permanently used by metadata. These
+ * blocks are generally:
+ * - the AG superblock, AGF, AGI and AGFL
+ * - the AGF (bno and cnt) and AGI btree root blocks, and optionally
+ * the AGI free inode and rmap btree root blocks.
+ * - blocks on the AGFL according to xfs_alloc_set_aside() limits
+ *
+ * The AG headers are sector sized, so the amount of space they take up is
+ * dependent on filesystem geometry. The others are all single blocks.
+ */
+unsigned int
+xfs_alloc_ag_max_usable(struct xfs_mount *mp)
+{
+ unsigned int blocks;
+
+ blocks = XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)); /* ag headers */
+ blocks += XFS_ALLOC_AGFL_RESERVE;
+ blocks += 3; /* AGF, AGI btree root blocks */
+ if (xfs_sb_version_hasfinobt(&mp->m_sb))
+ blocks++; /* finobt root block */
+ if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
+ /* rmap root block + full tree split on full AG */
+ blocks += 1 + (2 * mp->m_ag_maxlevels) - 1;
+ }
+
+ return mp->m_sb.sb_agblocks - blocks;
+}
+
+/*
* Lookup the record equal to [bno, len] in the btree given by cur.
*/
STATIC int /* error */
@@ -1911,6 +1977,9 @@ xfs_alloc_min_freelist(
/* space needed by-size freespace btree */
min_free += min_t(unsigned int, pag->pagf_levels[XFS_BTNUM_CNTi] + 1,
mp->m_ag_maxlevels);
+ /* space needed reverse mapping used space btree */
+ min_free += min_t(unsigned int, pag->pagf_levels[XFS_BTNUM_RMAPi] + 1,
+ mp->m_ag_maxlevels);
return min_free;
}
diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h
index 95b161f..67e564e 100644
--- a/fs/xfs/libxfs/xfs_alloc.h
+++ b/fs/xfs/libxfs/xfs_alloc.h
@@ -56,42 +56,6 @@ typedef unsigned int xfs_alloctype_t;
#define XFS_ALLOC_FLAG_FREEING 0x00000002 /* indicate caller is freeing extents*/
/*
- * In order to avoid ENOSPC-related deadlock caused by
- * out-of-order locking of AGF buffer (PV 947395), we place
- * constraints on the relationship among actual allocations for
- * data blocks, freelist blocks, and potential file data bmap
- * btree blocks. However, these restrictions may result in no
- * actual space allocated for a delayed extent, for example, a data
- * block in a certain AG is allocated but there is no additional
- * block for the additional bmap btree block due to a split of the
- * bmap btree of the file. The result of this may lead to an
- * infinite loop in xfssyncd when the file gets flushed to disk and
- * all delayed extents need to be actually allocated. To get around
- * this, we explicitly set aside a few blocks which will not be
- * reserved in delayed allocation. Considering the minimum number of
- * needed freelist blocks is 4 fsbs _per AG_, a potential split of file's bmap
- * btree requires 1 fsb, so we set the number of set-aside blocks
- * to 4 + 4*agcount.
- */
-#define XFS_ALLOC_SET_ASIDE(mp) (4 + ((mp)->m_sb.sb_agcount * 4))
-
-/*
- * When deciding how much space to allocate out of an AG, we limit the
- * allocation maximum size to the size the AG. However, we cannot use all the
- * blocks in the AG - some are permanently used by metadata. These
- * blocks are generally:
- * - the AG superblock, AGF, AGI and AGFL
- * - the AGF (bno and cnt) and AGI btree root blocks
- * - 4 blocks on the AGFL according to XFS_ALLOC_SET_ASIDE() limits
- *
- * The AG headers are sector sized, so the amount of space they take up is
- * dependent on filesystem geometry. The others are all single blocks.
- */
-#define XFS_ALLOC_AG_MAX_USABLE(mp) \
- ((mp)->m_sb.sb_agblocks - XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)) - 7)
-
-
-/*
* Argument structure for xfs_alloc routines.
* This is turned into a structure to avoid having 20 arguments passed
* down several levels of the stack.
@@ -131,6 +95,11 @@ typedef struct xfs_alloc_arg {
#define XFS_ALLOC_USERDATA 1 /* allocation is for user data*/
#define XFS_ALLOC_INITIAL_USER_DATA 2 /* special case start of file */
+/* freespace limit calculations */
+#define XFS_ALLOC_AGFL_RESERVE 4
+unsigned int xfs_alloc_set_aside(struct xfs_mount *mp);
+unsigned int xfs_alloc_ag_max_usable(struct xfs_mount *mp);
+
xfs_extlen_t xfs_alloc_longest_free_extent(struct xfs_mount *mp,
struct xfs_perag *pag, xfs_extlen_t need);
unsigned int xfs_alloc_min_freelist(struct xfs_mount *mp,
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index fef3767..e740ef5 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -3716,7 +3716,7 @@ xfs_bmap_btalloc(
args.fsbno = ap->blkno;
/* Trim the allocation back to the maximum an AG can fit. */
- args.maxlen = MIN(ap->length, XFS_ALLOC_AG_MAX_USABLE(mp));
+ args.maxlen = MIN(ap->length, mp->m_ag_max_usable);
args.firstblock = *ap->firstblock;
blen = 0;
if (nullfb) {
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 42b5696..0103a75 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -732,6 +732,8 @@ xfs_sb_mount_common(
mp->m_ialloc_min_blks = sbp->sb_spino_align;
else
mp->m_ialloc_min_blks = mp->m_ialloc_blks;
+ mp->m_alloc_set_aside = xfs_alloc_set_aside(mp);
+ mp->m_ag_max_usable = xfs_alloc_ag_max_usable(mp);
}
/*
diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
index e85a951..ec7bb8b 100644
--- a/fs/xfs/xfs_discard.c
+++ b/fs/xfs/xfs_discard.c
@@ -179,7 +179,7 @@ xfs_ioc_trim(
* matter as trimming blocks is an advisory interface.
*/
if (range.start >= XFS_FSB_TO_B(mp, mp->m_sb.sb_dblocks) ||
- range.minlen > XFS_FSB_TO_B(mp, XFS_ALLOC_AG_MAX_USABLE(mp)) ||
+ range.minlen > XFS_FSB_TO_B(mp, mp->m_ag_max_usable) ||
range.len < mp->m_sb.sb_blocksize)
return -EINVAL;
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 119be0a..031dd92 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -723,7 +723,7 @@ xfs_fs_counts(
cnt->allocino = percpu_counter_read_positive(&mp->m_icount);
cnt->freeino = percpu_counter_read_positive(&mp->m_ifree);
cnt->freedata = percpu_counter_read_positive(&mp->m_fdblocks) -
- XFS_ALLOC_SET_ASIDE(mp);
+ mp->m_alloc_set_aside;
spin_lock(&mp->m_sb_lock);
cnt->freertx = mp->m_sb.sb_frextents;
@@ -796,7 +796,7 @@ retry:
__int64_t free;
free = percpu_counter_sum(&mp->m_fdblocks) -
- XFS_ALLOC_SET_ASIDE(mp);
+ mp->m_alloc_set_aside;
if (!free)
goto out; /* ENOSPC and fdblks_delta = 0 */
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 1b2f72c..88da908 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -1196,7 +1196,7 @@ xfs_mod_fdblocks(
batch = XFS_FDBLOCKS_BATCH;
__percpu_counter_add(&mp->m_fdblocks, delta, batch);
- if (__percpu_counter_compare(&mp->m_fdblocks, XFS_ALLOC_SET_ASIDE(mp),
+ if (__percpu_counter_compare(&mp->m_fdblocks, mp->m_alloc_set_aside,
XFS_FDBLOCKS_BATCH) >= 0) {
/* we had space! */
return 0;
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 8030627..cdced0b 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -96,6 +96,8 @@ typedef struct xfs_mount {
uint m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */
uint m_in_maxlevels; /* max inobt btree levels. */
xfs_extlen_t m_ag_prealloc_blocks; /* reserved ag blocks */
+ uint m_alloc_set_aside; /* space we can't use */
+ uint m_ag_max_usable; /* max space per AG */
struct radix_tree_root m_perag_tree; /* per-ag accounting info */
spinlock_t m_perag_lock; /* lock for m_perag_tree */
struct mutex m_growlock; /* growfs mutex */
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 904f637..6dc5993 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -1072,7 +1072,7 @@ xfs_fs_statfs(
statp->f_blocks = sbp->sb_dblocks - lsize;
spin_unlock(&mp->m_sb_lock);
- statp->f_bfree = fdblocks - XFS_ALLOC_SET_ASIDE(mp);
+ statp->f_bfree = fdblocks - mp->m_alloc_set_aside;
statp->f_bavail = statp->f_bfree;
fakeinos = statp->f_bfree << sbp->sb_inopblog;
WARNING: multiple messages have this Message-ID (diff)
From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: david@fromorbit.com, darrick.wong@oracle.com
Cc: linux-fsdevel@vger.kernel.org, Dave Chinner <dchinner@redhat.com>,
xfs@oss.sgi.com
Subject: [PATCH 15/58] xfs: rmap btree requires more reserved free space
Date: Tue, 06 Oct 2015 21:56:25 -0700 [thread overview]
Message-ID: <20151007045625.30457.21863.stgit@birch.djwong.org> (raw)
In-Reply-To: <20151007045443.30457.47038.stgit@birch.djwong.org>
>From : Dave Chinner <dchinner@redhat.com>
The rmap btree is allocated from the AGFL, which means we have to
ensure ENOSPC is reported to userspace before we run out of free
space in each AG. The last allocation in an AG can cause a full
height rmap btree split, and that means we have to reserve at least
this many blocks *in each AG* to be placed on the AGFL at ENOSPC.
Update the various space calculation functiosn to handle this.
Also, because the macros are now executing conditional code and are called quite
frequently, convert them to functions that initialise varaibles in the struct
xfs_mount, use the new variables everywhere and document the calculations
better.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
---
fs/xfs/libxfs/xfs_alloc.c | 69 +++++++++++++++++++++++++++++++++++++++++++++
fs/xfs/libxfs/xfs_alloc.h | 41 +++------------------------
fs/xfs/libxfs/xfs_bmap.c | 2 +
fs/xfs/libxfs/xfs_sb.c | 2 +
fs/xfs/xfs_discard.c | 2 +
fs/xfs/xfs_fsops.c | 4 +--
fs/xfs/xfs_mount.c | 2 +
fs/xfs/xfs_mount.h | 2 +
fs/xfs/xfs_super.c | 2 +
9 files changed, 84 insertions(+), 42 deletions(-)
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index 3c55fa7..6fd9d3e 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -62,6 +62,72 @@ xfs_prealloc_blocks(
}
/*
+ * In order to avoid ENOSPC-related deadlock caused by out-of-order locking of
+ * AGF buffer (PV 947395), we place constraints on the relationship among actual
+ * allocations for data blocks, freelist blocks, and potential file data bmap
+ * btree blocks. However, these restrictions may result in no actual space
+ * allocated for a delayed extent, for example, a data block in a certain AG is
+ * allocated but there is no additional block for the additional bmap btree
+ * block due to a split of the bmap btree of the file. The result of this may
+ * lead to an infinite loop when the file gets flushed to disk and all delayed
+ * extents need to be actually allocated. To get around this, we explicitly set
+ * aside a few blocks which will not be reserved in delayed allocation.
+ *
+ * The minimum number of needed freelist blocks is 4 fsbs _per AG_ when we are
+ * not using rmap btrees a potential split of file's bmap btree requires 1 fsb,
+ * so we set the number of set-aside blocks to 4 + 4*agcount when not using rmap
+ * btrees.
+ *
+ * When rmap btrees are active, we have to consider that using the last block in
+ * the AG can cause a full height rmap btree split and we need enough blocks on
+ * the AGFL to be able to handle this. That means we have, in addition to the
+ * above consideration, another (2 * mp->m_ag_levels) - 1 blocks required to be
+ * available to the free list.
+ */
+unsigned int
+xfs_alloc_set_aside(
+ struct xfs_mount *mp)
+{
+ unsigned int blocks;
+
+ blocks = 4 + (mp->m_sb.sb_agcount * XFS_ALLOC_AGFL_RESERVE);
+ if (!xfs_sb_version_hasrmapbt(&mp->m_sb))
+ return blocks;
+ return blocks + (mp->m_sb.sb_agcount * (2 * mp->m_ag_maxlevels) - 1);
+}
+
+/*
+ * When deciding how much space to allocate out of an AG, we limit the
+ * allocation maximum size to the size the AG. However, we cannot use all the
+ * blocks in the AG - some are permanently used by metadata. These
+ * blocks are generally:
+ * - the AG superblock, AGF, AGI and AGFL
+ * - the AGF (bno and cnt) and AGI btree root blocks, and optionally
+ * the AGI free inode and rmap btree root blocks.
+ * - blocks on the AGFL according to xfs_alloc_set_aside() limits
+ *
+ * The AG headers are sector sized, so the amount of space they take up is
+ * dependent on filesystem geometry. The others are all single blocks.
+ */
+unsigned int
+xfs_alloc_ag_max_usable(struct xfs_mount *mp)
+{
+ unsigned int blocks;
+
+ blocks = XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)); /* ag headers */
+ blocks += XFS_ALLOC_AGFL_RESERVE;
+ blocks += 3; /* AGF, AGI btree root blocks */
+ if (xfs_sb_version_hasfinobt(&mp->m_sb))
+ blocks++; /* finobt root block */
+ if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
+ /* rmap root block + full tree split on full AG */
+ blocks += 1 + (2 * mp->m_ag_maxlevels) - 1;
+ }
+
+ return mp->m_sb.sb_agblocks - blocks;
+}
+
+/*
* Lookup the record equal to [bno, len] in the btree given by cur.
*/
STATIC int /* error */
@@ -1911,6 +1977,9 @@ xfs_alloc_min_freelist(
/* space needed by-size freespace btree */
min_free += min_t(unsigned int, pag->pagf_levels[XFS_BTNUM_CNTi] + 1,
mp->m_ag_maxlevels);
+ /* space needed reverse mapping used space btree */
+ min_free += min_t(unsigned int, pag->pagf_levels[XFS_BTNUM_RMAPi] + 1,
+ mp->m_ag_maxlevels);
return min_free;
}
diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h
index 95b161f..67e564e 100644
--- a/fs/xfs/libxfs/xfs_alloc.h
+++ b/fs/xfs/libxfs/xfs_alloc.h
@@ -56,42 +56,6 @@ typedef unsigned int xfs_alloctype_t;
#define XFS_ALLOC_FLAG_FREEING 0x00000002 /* indicate caller is freeing extents*/
/*
- * In order to avoid ENOSPC-related deadlock caused by
- * out-of-order locking of AGF buffer (PV 947395), we place
- * constraints on the relationship among actual allocations for
- * data blocks, freelist blocks, and potential file data bmap
- * btree blocks. However, these restrictions may result in no
- * actual space allocated for a delayed extent, for example, a data
- * block in a certain AG is allocated but there is no additional
- * block for the additional bmap btree block due to a split of the
- * bmap btree of the file. The result of this may lead to an
- * infinite loop in xfssyncd when the file gets flushed to disk and
- * all delayed extents need to be actually allocated. To get around
- * this, we explicitly set aside a few blocks which will not be
- * reserved in delayed allocation. Considering the minimum number of
- * needed freelist blocks is 4 fsbs _per AG_, a potential split of file's bmap
- * btree requires 1 fsb, so we set the number of set-aside blocks
- * to 4 + 4*agcount.
- */
-#define XFS_ALLOC_SET_ASIDE(mp) (4 + ((mp)->m_sb.sb_agcount * 4))
-
-/*
- * When deciding how much space to allocate out of an AG, we limit the
- * allocation maximum size to the size the AG. However, we cannot use all the
- * blocks in the AG - some are permanently used by metadata. These
- * blocks are generally:
- * - the AG superblock, AGF, AGI and AGFL
- * - the AGF (bno and cnt) and AGI btree root blocks
- * - 4 blocks on the AGFL according to XFS_ALLOC_SET_ASIDE() limits
- *
- * The AG headers are sector sized, so the amount of space they take up is
- * dependent on filesystem geometry. The others are all single blocks.
- */
-#define XFS_ALLOC_AG_MAX_USABLE(mp) \
- ((mp)->m_sb.sb_agblocks - XFS_BB_TO_FSB(mp, XFS_FSS_TO_BB(mp, 4)) - 7)
-
-
-/*
* Argument structure for xfs_alloc routines.
* This is turned into a structure to avoid having 20 arguments passed
* down several levels of the stack.
@@ -131,6 +95,11 @@ typedef struct xfs_alloc_arg {
#define XFS_ALLOC_USERDATA 1 /* allocation is for user data*/
#define XFS_ALLOC_INITIAL_USER_DATA 2 /* special case start of file */
+/* freespace limit calculations */
+#define XFS_ALLOC_AGFL_RESERVE 4
+unsigned int xfs_alloc_set_aside(struct xfs_mount *mp);
+unsigned int xfs_alloc_ag_max_usable(struct xfs_mount *mp);
+
xfs_extlen_t xfs_alloc_longest_free_extent(struct xfs_mount *mp,
struct xfs_perag *pag, xfs_extlen_t need);
unsigned int xfs_alloc_min_freelist(struct xfs_mount *mp,
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index fef3767..e740ef5 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -3716,7 +3716,7 @@ xfs_bmap_btalloc(
args.fsbno = ap->blkno;
/* Trim the allocation back to the maximum an AG can fit. */
- args.maxlen = MIN(ap->length, XFS_ALLOC_AG_MAX_USABLE(mp));
+ args.maxlen = MIN(ap->length, mp->m_ag_max_usable);
args.firstblock = *ap->firstblock;
blen = 0;
if (nullfb) {
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 42b5696..0103a75 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -732,6 +732,8 @@ xfs_sb_mount_common(
mp->m_ialloc_min_blks = sbp->sb_spino_align;
else
mp->m_ialloc_min_blks = mp->m_ialloc_blks;
+ mp->m_alloc_set_aside = xfs_alloc_set_aside(mp);
+ mp->m_ag_max_usable = xfs_alloc_ag_max_usable(mp);
}
/*
diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
index e85a951..ec7bb8b 100644
--- a/fs/xfs/xfs_discard.c
+++ b/fs/xfs/xfs_discard.c
@@ -179,7 +179,7 @@ xfs_ioc_trim(
* matter as trimming blocks is an advisory interface.
*/
if (range.start >= XFS_FSB_TO_B(mp, mp->m_sb.sb_dblocks) ||
- range.minlen > XFS_FSB_TO_B(mp, XFS_ALLOC_AG_MAX_USABLE(mp)) ||
+ range.minlen > XFS_FSB_TO_B(mp, mp->m_ag_max_usable) ||
range.len < mp->m_sb.sb_blocksize)
return -EINVAL;
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 119be0a..031dd92 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -723,7 +723,7 @@ xfs_fs_counts(
cnt->allocino = percpu_counter_read_positive(&mp->m_icount);
cnt->freeino = percpu_counter_read_positive(&mp->m_ifree);
cnt->freedata = percpu_counter_read_positive(&mp->m_fdblocks) -
- XFS_ALLOC_SET_ASIDE(mp);
+ mp->m_alloc_set_aside;
spin_lock(&mp->m_sb_lock);
cnt->freertx = mp->m_sb.sb_frextents;
@@ -796,7 +796,7 @@ retry:
__int64_t free;
free = percpu_counter_sum(&mp->m_fdblocks) -
- XFS_ALLOC_SET_ASIDE(mp);
+ mp->m_alloc_set_aside;
if (!free)
goto out; /* ENOSPC and fdblks_delta = 0 */
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 1b2f72c..88da908 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -1196,7 +1196,7 @@ xfs_mod_fdblocks(
batch = XFS_FDBLOCKS_BATCH;
__percpu_counter_add(&mp->m_fdblocks, delta, batch);
- if (__percpu_counter_compare(&mp->m_fdblocks, XFS_ALLOC_SET_ASIDE(mp),
+ if (__percpu_counter_compare(&mp->m_fdblocks, mp->m_alloc_set_aside,
XFS_FDBLOCKS_BATCH) >= 0) {
/* we had space! */
return 0;
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 8030627..cdced0b 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -96,6 +96,8 @@ typedef struct xfs_mount {
uint m_bm_maxlevels[2]; /* XFS_BM_MAXLEVELS */
uint m_in_maxlevels; /* max inobt btree levels. */
xfs_extlen_t m_ag_prealloc_blocks; /* reserved ag blocks */
+ uint m_alloc_set_aside; /* space we can't use */
+ uint m_ag_max_usable; /* max space per AG */
struct radix_tree_root m_perag_tree; /* per-ag accounting info */
spinlock_t m_perag_lock; /* lock for m_perag_tree */
struct mutex m_growlock; /* growfs mutex */
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 904f637..6dc5993 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -1072,7 +1072,7 @@ xfs_fs_statfs(
statp->f_blocks = sbp->sb_dblocks - lsize;
spin_unlock(&mp->m_sb_lock);
- statp->f_bfree = fdblocks - XFS_ALLOC_SET_ASIDE(mp);
+ statp->f_bfree = fdblocks - mp->m_alloc_set_aside;
statp->f_bavail = statp->f_bfree;
fakeinos = statp->f_bfree << sbp->sb_inopblog;
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2015-10-07 4:56 UTC|newest]
Thread overview: 131+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-07 4:54 [RFCv3 00/58] xfs: add reverse-mapping, reflink, and dedupe support Darrick J. Wong
2015-10-07 4:54 ` Darrick J. Wong
2015-10-07 4:54 ` [PATCH 01/58] libxfs: make xfs_alloc_fix_freelist non-static Darrick J. Wong
2015-10-07 4:54 ` Darrick J. Wong
2015-10-07 4:54 ` [PATCH 02/58] xfs: fix log ticket type printing Darrick J. Wong
2015-10-07 4:54 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 03/58] xfs: introduce rmap btree definitions Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 04/58] xfs: add rmap btree stats infrastructure Darrick J. Wong
2015-10-07 4:55 ` [PATCH 05/58] xfs: rmap btree add more reserved blocks Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 06/58] xfs: add owner field to extent allocation and freeing Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 07/58] xfs: add extended " Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 08/58] xfs: introduce rmap extent operation stubs Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 09/58] xfs: extend rmap extent operation stubs to take full owner info Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 10/58] xfs: define the on-disk rmap btree format Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:55 ` [PATCH 11/58] xfs: enhance " Darrick J. Wong
2015-10-07 4:55 ` Darrick J. Wong
2015-10-07 4:56 ` [PATCH 12/58] xfs: add rmap btree growfs support Darrick J. Wong
2015-10-07 4:56 ` Darrick J. Wong
2015-10-07 4:56 ` [PATCH 13/58] xfs: enhance " Darrick J. Wong
2015-10-07 4:56 ` Darrick J. Wong
2015-10-07 4:56 ` [PATCH 14/58] xfs: rmap btree transaction reservations Darrick J. Wong
2015-10-07 4:56 ` Darrick J. Wong
2015-10-07 4:56 ` Darrick J. Wong [this message]
2015-10-07 4:56 ` [PATCH 15/58] xfs: rmap btree requires more reserved free space Darrick J. Wong
2015-10-07 4:56 ` [PATCH 16/58] libxfs: fix min freelist length calculation Darrick J. Wong
2015-10-07 4:56 ` Darrick J. Wong
2015-10-07 4:56 ` [PATCH 17/58] xfs: add rmap btree operations Darrick J. Wong
2015-10-07 4:57 ` [PATCH 18/58] xfs: enhance " Darrick J. Wong
2015-10-07 4:57 ` [PATCH 19/58] xfs: add an extent to the rmap btree Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-07 4:57 ` [PATCH 20/58] xfs: add tracepoints for the rmap-mirrors-bmbt functions Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-07 4:57 ` [PATCH 21/58] xfs: teach rmap_alloc how to deal with our larger rmap btree Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-07 4:57 ` [PATCH 22/58] xfs: remove an extent from the " Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-07 4:57 ` [PATCH 23/58] xfs: enhanced " Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-07 4:57 ` [PATCH 24/58] xfs: add rmap btree insert and delete helpers Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-07 4:57 ` [PATCH 25/58] xfs: bmap btree changes should update rmap btree Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-21 21:39 ` Darrick J. Wong
2015-10-21 21:39 ` Darrick J. Wong
2015-10-07 4:57 ` [PATCH 26/58] xfs: add rmap btree geometry feature flag Darrick J. Wong
2015-10-07 4:57 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 27/58] xfs: add rmap btree block detection to log recovery Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 28/58] xfs: enable the rmap btree functionality Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 29/58] xfs: disable XFS_IOC_SWAPEXT when rmap btree is enabled Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 30/58] xfs: implement " Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 31/58] libxfs: refactor short btree block verification Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 32/58] xfs: don't update rmapbt when fixing agfl Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 33/58] xfs: introduce refcount btree definitions Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 34/58] xfs: add refcount btree stats infrastructure Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:58 ` [PATCH 35/58] xfs: refcount btree add more reserved blocks Darrick J. Wong
2015-10-07 4:58 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 36/58] xfs: define the on-disk refcount btree format Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 37/58] xfs: define tracepoints for refcount/reflink activities Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 38/58] xfs: add refcount btree support to growfs Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 39/58] xfs: add refcount btree operations Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 40/58] libxfs: adjust refcount of an extent of blocks in refcount btree Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-27 19:05 ` Darrick J. Wong
2015-10-27 19:05 ` Darrick J. Wong
2015-10-30 20:56 ` Darrick J. Wong
2015-10-30 20:56 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 41/58] libxfs: adjust refcount when unmapping file blocks Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 42/58] xfs: add refcount btree block detection to log recovery Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 43/58] xfs: map an inode's offset to an exact physical block Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 4:59 ` [PATCH 44/58] xfs: add reflink feature flag to geometry Darrick J. Wong
2015-10-07 4:59 ` Darrick J. Wong
2015-10-07 5:00 ` [PATCH 45/58] xfs: create a separate workqueue for copy-on-write activities Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:00 ` [PATCH 46/58] xfs: implement copy-on-write for reflinked blocks Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:00 ` [PATCH 47/58] xfs: handle directio " Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:00 ` [PATCH 48/58] xfs: copy-on-write reflinked blocks when zeroing ranges of blocks Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-21 21:17 ` Darrick J. Wong
2015-10-21 21:17 ` Darrick J. Wong
2015-10-07 5:00 ` [PATCH 49/58] xfs: clear inode reflink flag when freeing blocks Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:00 ` [PATCH 50/58] xfs: reflink extents from one file to another Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:12 ` kbuild test robot
2015-10-07 5:12 ` kbuild test robot
2015-10-07 5:00 ` [PATCH 51/58] xfs: add clone file and clone range ioctls Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:13 ` kbuild test robot
2015-10-07 5:13 ` kbuild test robot
2015-10-07 6:46 ` kbuild test robot
2015-10-07 6:46 ` kbuild test robot
2015-10-07 7:35 ` kbuild test robot
2015-10-07 7:35 ` kbuild test robot
2015-10-07 5:00 ` [PATCH 52/58] xfs: emulate the btrfs dedupe extent same ioctl Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:00 ` [PATCH 53/58] xfs: teach fiemap about reflink'd extents Darrick J. Wong
2015-10-07 5:00 ` Darrick J. Wong
2015-10-07 5:01 ` [PATCH 54/58] xfs: swap inode reflink flags when swapping inode extents Darrick J. Wong
2015-10-07 5:01 ` Darrick J. Wong
2015-10-07 5:01 ` [PATCH 55/58] vfs: add a FALLOC_FL_UNSHARE mode to fallocate to unshare a range of blocks Darrick J. Wong
2015-10-07 5:01 ` Darrick J. Wong
2015-10-07 5:01 ` [PATCH 56/58] xfs: unshare a range of blocks via fallocate Darrick J. Wong
2015-10-07 5:01 ` Darrick J. Wong
2015-10-07 5:01 ` [PATCH 57/58] xfs: support XFS_XFLAG_REFLINK (and FS_NOCOW_FL) on reflink filesystems Darrick J. Wong
2015-10-07 5:01 ` Darrick J. Wong
2015-10-07 5:01 ` [PATCH 58/58] xfs: recognize the reflink feature bit Darrick J. Wong
2015-10-07 5:01 ` Darrick J. Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20151007045625.30457.21863.stgit@birch.djwong.org \
--to=darrick.wong@oracle.com \
--cc=david@fromorbit.com \
--cc=dchinner@redhat.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=xfs@oss.sgi.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.